[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ZNZq76jCnzNy7QVF@google.com>
Date: Fri, 11 Aug 2023 10:07:59 -0700
From: Stanislav Fomichev <sdf@...gle.com>
To: Rong Tao <rtoax@...mail.com>
Cc: ast@...nel.org, rongtao@...tc.cn,
Andrii Nakryiko <andrii@...nel.org>,
Mykola Lysenko <mykolal@...com>,
Daniel Borkmann <daniel@...earbox.net>,
Martin KaFai Lau <martin.lau@...ux.dev>,
Song Liu <song@...nel.org>,
Yonghong Song <yonghong.song@...ux.dev>,
John Fastabend <john.fastabend@...il.com>,
KP Singh <kpsingh@...nel.org>, Hao Luo <haoluo@...gle.com>,
Jiri Olsa <jolsa@...nel.org>, Shuah Khan <shuah@...nel.org>,
"open list:BPF [SELFTESTS] (Test Runners & Infrastructure)"
<bpf@...r.kernel.org>,
"open list:KERNEL SELFTEST FRAMEWORK"
<linux-kselftest@...r.kernel.org>,
open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH bpf-next v2] selftests/bpf: trace_helpers.c: optimize
kallsyms cache
On 08/11, Rong Tao wrote:
> From: Rong Tao <rongtao@...tc.cn>
>
> Static ksyms often have problems because the number of symbols exceeds the
> MAX_SYMS limit. Like changing the MAX_SYMS from 300000 to 400000 in
> commit e76a014334a6("selftests/bpf: Bump and validate MAX_SYMS") solves
> the problem somewhat, but it's not the perfect way.
>
> This commit uses dynamic memory allocation, which completely solves the
> problem caused by the limitation of the number of kallsyms.
>
> Signed-off-by: Rong Tao <rongtao@...tc.cn>
> ---
> v2: Do the usual len/capacity scheme here to amortize the cost of realloc, and
> don't free symbols.
> v1: https://lore.kernel.org/lkml/tencent_AB461510B10CD484E0B2F62E3754165F2909@qq.com/
> ---
> tools/testing/selftests/bpf/trace_helpers.c | 73 ++++++++++++++-------
> 1 file changed, 48 insertions(+), 25 deletions(-)
>
> diff --git a/tools/testing/selftests/bpf/trace_helpers.c b/tools/testing/selftests/bpf/trace_helpers.c
> index f83d9f65c65b..cda5a2328450 100644
> --- a/tools/testing/selftests/bpf/trace_helpers.c
> +++ b/tools/testing/selftests/bpf/trace_helpers.c
> @@ -18,9 +18,37 @@
> #define TRACEFS_PIPE "/sys/kernel/tracing/trace_pipe"
> #define DEBUGFS_PIPE "/sys/kernel/debug/tracing/trace_pipe"
>
> -#define MAX_SYMS 400000
> -static struct ksym syms[MAX_SYMS];
> -static int sym_cnt;
> +static struct {
> + struct ksym *syms;
> + unsigned int sym_cap;
> + unsigned int sym_cnt;
> +} ksyms = {
> + .syms = NULL,
> + .sym_cap = 1024,
> + .sym_cnt = 0,
> +};
Not sure what the struct buys you here (besides grouping everything
nicely), maybe do the following?
static struct ksym *syms;
static int sym_cnt;
static int sym_cap = 1024;
Will reduce the churn elsewhere..
> +static int ksyms__add_symbol(const char *name, unsigned long addr)
> +{
> + void *tmp;
> + unsigned int new_cap;
> +
> + if (ksyms.sym_cnt + 1 > ksyms.sym_cap) {
> + new_cap = ksyms.sym_cap * 4 / 3;
> + tmp = realloc(ksyms.syms, sizeof(struct ksym) * new_cap);
> + if (!tmp)
> + return -ENOMEM;
> + ksyms.syms = tmp;
> + ksyms.sym_cap = new_cap;
> + }
> +
> + ksyms.syms[ksyms.sym_cnt].addr = addr;
> + ksyms.syms[ksyms.sym_cnt].name = strdup(name);
> +
> + ksyms.sym_cnt++;
> +
> + return 0;
> +}
>
> static int ksym_cmp(const void *p1, const void *p2)
> {
> @@ -33,9 +61,10 @@ int load_kallsyms_refresh(void)
> char func[256], buf[256];
> char symbol;
> void *addr;
> - int i = 0;
>
> - sym_cnt = 0;
> + ksyms.syms = malloc(sizeof(struct ksym) * ksyms.sym_cap);
> + if (!ksyms.syms)
> + return -ENOMEM;
>
> f = fopen("/proc/kallsyms", "r");
> if (!f)
> @@ -46,16 +75,10 @@ int load_kallsyms_refresh(void)
> break;
> if (!addr)
> continue;
> - if (i >= MAX_SYMS)
> - return -EFBIG;
> -
> - syms[i].addr = (long) addr;
> - syms[i].name = strdup(func);
> - i++;
> + ksyms__add_symbol(func, (unsigned long)addr);
Need to check the return of ksyms__add_symbol here?
Powered by blists - more mailing lists